As the AI Impact Summit unfolds in New Delhi, India finds itself at a crossroads: should it create its own foundational AI models, or continue building on global platforms dominated by US and Chinese giants? Outside Bharat Mandapam — the summit’s venue — the city has been polished for the occasion. Inside, the debate is far more complex, reported Bloomberg.
India’s AI story is already showing signs of both promise and pressure.
Fractal Analytics, India’s first AI unicorn, slipped 5% on its stock market debut days after raising $313 million in an IPO. At the same time, a Blackstone-led investor group is planning to invest up to $600 million in Indian AI cloud startup Neysa.
The contrasting signals go beyond public versus private market risk appetite. They reflect deeper tensions — limited compute power, crowded AI services, and the uncertain pace of monetization. The AI cycle in India will likely swing between euphoria and correction many times over.
Sovereign Models vs Global Giants
The global AI landscape is dominated by OpenAI’s GPT, Anthropic’s Claude and Google’s Gemini. China has DeepSeek. India, so far, has no model at that scale.
That may begin to change. BharatGen and Sarvam AI are set to launch sovereign large language models tailored to Indian languages and voice-first use cases. Several specialized models serving narrower business needs are already operational.
Sarvam AI co-founder Vivek Raghavan has argued that without building domestic capabilities, India risks becoming a “digital tenant” in a foreign ecosystem. Leveraging frontier models is necessary, he says, but developing sovereign alternatives is equally critical.
The underlying question: can India shape AI infrastructure, or will it merely consume it?
The global shift from generative AI — building presentations and drafting emails — to agentic AI — managing supply chains and automating workflows — is underway.
Source: India Latest News, Breaking News Today, Top News Headlines | Times Now
